A review and meta-analysis of 42 group and ringlecase studies evaluating antecedent exercise (AE) as a means of reducing disrupti-behaviors was conducted.Of 16 group studies, 12 produced positive results and 4 produced negative results. The weighted mean effect size, expressed as Cohen's d, was .33 with a standard error of .08. Moderator analysis indicated that studies using direct behavioral obsewation, hyperactive subjects, or nonaerobic exercise obtained greater effects, and that studies of higher quality obtained weaker effects. Of 26 single-use studies, 22 produced positive msults, 1 produced no effect, and 3 produced negative results. The weighted mean effect size, expressed as d, was 1.99 with a standard error of .411. Among the single-subject studies, moderator analyses were unable t o detect statistically significant moderatom of effect size. Information was reviewed suggesting that AE is socially acceptable, can be implemented with treatment integrity, and has a benign side effoct profile. The extent t o which AE is functionally based remains open to question due to a lack of understanding regarding mechanism of action. Ten hypothesized mechanisms of action are briefly discussed.